April 23, 2014

Lochsbour's IBD in modern Europeans is greatest among Danish but most direct among French

This is a most interesting issue I forgot to discuss when previously addressing the massively interesting Lazaridis et al. study on European ancestry based on ancient autosomal DNA (see here and here). 

Identity by descent (IBD) data shows interesting differences between populations in Supplemental Information's article 18. While Stuttgart's (early farmer) ancestry is more or less the same by both measures (Sardinians first, followed by Slovakians and some other Balcanic and Central European populations), there are important differences in the ancestry of Lochsbour (Epipaleolithic hunter-gatherer from Luxembourg). While the Danes score highest in overall IBD block number (rough relatedness to Lochsbour) it is the French who score highest in IBD length (indicating a more direct relatedness, even if in smaller amounts). 

The difference between the French and Danish is quite significant, I believe, and seems to suggest that Lochsbour's relatives had a direct impact on modern French genetics, while the impact of Lochsbour as such on other populations should be considered more indirect (i.e. via other hunter-gatherer populations). 

This implies that there was some important diversity among the hunter-gatherer groups that influenced modern European genetics and that Lochsbour must be considered a mere generic proxy. Possibly if Motala or La Braña would have been used as reference instead, we would get some important differences in the results, as would be the case if Balcanic or Eastern European hunter-gatherers would be thrown into the equation, no doubt. 

You may have noticed that there are some notable samples unmarked in the graphs, that's because they are colonial populations such as Zimbabwean or North American Whites, whose exact ancestry is not easy to track. The green and red texts are my illustrative additions.

While not marked, I find also notable and rather perplexing that Lebanon shows up as the fourth non-colonial population more related to Lochsbour by IBD length, after Scotland but before Ukraine, the Netherlands and Sweden.

In any case you can parse the data for the 10 more notable samples of each measure in the supplemental material, chapter 18.

Referenced study:

Iosif Lazaridis, Nick Patterson, Alissa Mittnik, et al., Ancient human genomes suggest three ancestral populations for present-day Europeans. BioArxiv 2013 (preprint). Freely accessibleLINK [doi:10.1101/001552]


  1. The Danish sample is limited to one person, and so is the Slovakian one. In fact, the Slovakian appears to be of Jewish or Roma origin.

    These samples come from POPRES. Whoever collected them had a very strange sampling strategy.

    1. Ah, thanks, that's an interesting reference to know. I found a bit frustrating that SI-18 is not somewhat more exhaustive, but guess it was not their priority.

      IBD in any case should give important information about actual relatedness to each of the ancient samples used as proxy or whatever sub-population they represent. This seems particularly more important re. hunter-gatherers but it may also be significant re. farmers, judging on the Portalón and Gök. distinct clustering relative to Lazaridis' EEF sample.

      Anyhow, regarding to the Slovakian, it does seem interesting in any case that the sequence of pops. with greatest IBD Stuttgart ancestry is full of Balcanic and Central European samples. This suggests that there is some blank spot regarding other populations (SW Europeans, non-Sardinian Italians) which score high for Stuttgart affinity in the formal tests but not in the IBD results. Something in need of revision, right?

    2. I don't really know what to make of these results, since the sample numbers are so erratic from different parts of Europe. This analysis and its presentation just seem like an afterthought.

      Someone should maybe try it again, but I'd say ChromoPainter/fineSTRUCTURE are better at this sort of stuff than BEAGLE/fastIBD, and the results from an analysis using the former are featured in this study. The problem is that, as far as I know, there's very little data offered from that run, without, for instance, the segment (or chunk) length stats, which would be very interesting. Maybe they're available on request?

      By the way, here's a PCA of the European samples from the POPRES dataset. Note the position of the lone Slovakian. I'd say he's in large part of Ashkenazi Jewish descent.


    3. The Chromopainter graph is in this very same chapter but I can't see anything on it, except what is already amplified and even that with difficulty. How is Chromopainter better for this particular case? Does it even count chunk length or just chunk number, what is much less interesting?

      Here we see that, regardless of who has the most chunks or the most combined Lochsbour relatedness in general (be them Danes, Lithuanians or Finns or CEU or whatever, not really that important), what is clear is that the directness of Lochsbour's ancestry is concentrated quite outstandingly among the French.

      And that is the really interesting matter, because it implies that the age of whatever relatedness with the Lochsbour clan of that particular population is much more recent, much more direct than the rest.

      As for the Slovakian sample, I'm guessing that he can't be Jewish, let alone Roma: that would not produce a recent admixture date (IBD length) like that, unless they are like the anomalous Lebanese sample (which I now suspect as Christian of French or Italian descent). The guy must be related to that Carpathian pastoralist minority so rich in mtDNA N1a of Neolithic ancestry. I can't recall the name of the group nor I can find a reference right now but the population was certainly discussed in some forums back in the day. They are probably a Rusyn subgroup but can't find any confirmation right now.

      If you have access to this 2009 paper (→ http://digitalcommons.wayne.edu/humbiol/vol81/iss1/4/) maybe you can find something more specific.

      There must be small pockets of more direct "Danubian" ancestry in the Carpathian mountains for all I know.

    4. → http://www.biomedcentral.com/1471-2148/10/304

      Not the study I was looking for but anyhow it detects relatively high frequencies of an N1a subclade among Slovaks, related to some of the ancient Danubian farmers, as well as among other modern populations (different subclades).


Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (sorry, too many trolls).